Every year thousands of Americans die or experience serious complications undergoing elective cancer surgery. Wide variations in mortality rates across both hospitals and surgeons suggest that the safety of cancer surgery could be improved substantially, but quality improvement efforts are currently limited by a lack of understanding about mechanisms underlying variations in hospital performance. For example, are high mortality rates at these hospitals attributable to surgical or medical complications? Do high mortality hospitals simply have higher complication rates, or are they instead less proficient at managing complications once they have occurred? Answers to such questions have obvious implications for the types of interventions likely to effect meaningful quality improvement in cancer surgery. Evidence that excess mortality is attributable primarily to surgical site complications would imply the need for interventions aimed at ensuring surgeon proficiency. Variation in specific medical complications would instead suggest the need to focus on processes related to perioperative care. In this context, our project has two specific aims: 1. To determine the causes of excess operative deaths at hospitals with high cancer surgery mortality. Using data from the 2005-6 National Cancer Database, we will identify 20 hospitals with among the lowest cancer surgery mortality rates in the United States (approximately 1.5%) and 30 hospitals with the highest mortality rates (approximately 10%). Based on clinical chart review, we will first compare these two groups of hospitals with regard to cause-specific mortality rates. We will then examine whether differences in cause- specific mortality rates are attributable to differences in the incidence of complications or failure to rescue rates. 2. To identify resources and processes of care that account for differences in complication rates and mortality. With a better understanding of the clinical causes of excess deaths at high mortality hospitals, we will then examine specific structural variables and processes of care that underlie differences in outcomes between the 2 hospital groups. These include factors that may reduce the incidence of complications (e.g.., adherence to evidence-based practices in perioperative care) or improve rescue rates in patients with complications.Project Narrative: Although complication and mortality rates with cancer surgery vary widely across hospitals, efforts aimed improving surgical quality are hindered by a lack of understanding about exactly why some hospitals have better outcomes than others. In this national study of 50 hospitals, we will first determine the clinical causes of excess operative deaths at hospitals with high mortality. We will then identify the resources and processes of care that account for differences in hospital outcomes, with the ultimate goal of identifying best practices and improving care in all settings.